ACE: Cooperative Multi-agent Q-learning with Bidirectional Action-Dependency
Open paper
Benchmark
Model
Rank
Results
smac-on-smac-3s5z-vs-3s6z
ACE
#1
Median Win Rate: 100
smac-on-smac-6h-vs-8z
ACE
#1
Median Win Rate: 93.75
smac-on-smac-corridor
ACE
#1
Median Win Rate: 100
smac-on-smac-mmm2
ACE
#1
Median Win Rate: 100
Rank counts only results with a code link.